A Hybrid Solution to the Human/Machine Inspection Systems Design Problem in Manufacturing and Service Industries

摘要

Product inspection is an important step in ensuring product quality with one of its most important tasks being visual inspection. Because of its importance in ensuring quality, any improvement in inspection performance holds great potential for increasing an industry's competitiveness. If an inspection is to be successful, it is critical that the various functions constituting an inspection task be performed optimally. An inspection task typically consists of the following functions: orientation, search, decision-making, and recording. Orientation and recording, essentially manual activities, are best automated. The search and decision-making functions, however, are essentially cognitive activities and have been shown to be the most important determinants of inspection performance that system designers need guidance in allocating. With the customer demand for zero defects in products, 100% inspection using automated systems has seen more frequent application than traditional sampling inspection using human inspectors. Despite the advantages of automation, these inspection systems often fail to meet expectations primarily because they ignore humans' ability in pattern recognition, as rational decision-makers and their flexibility to adapt to new situations. Humans and computers each have advantages and disadvantages that system designers need to consider. For example, humans are good at detecting signals in overlapping noise spectra and can make inductive decisions in new situations; however, they are limited in their computational ability and short-term memory. On the other hand, computers are good at computation, memory storage and retrieving, but are poor at detecting signals in noise and have very little capacity for creative or inductive functions. Thus, designers of systems which include an inspection component need guidance on human/machine function allocation to ensure that the inspection is performed at the very least adequately and, preferably, effectively and efficiently. In response to this need, this paper describes a system that will facilitate the conducting of controlled studies to address issues related to human machine system design and function allocation in visual inspection. The system simulates the search and decision making functions of a Printed Circuit Board (PCB) inspection task. The system can operate in three separate modes: (1) human inspection mode - where all the functions are performed by the human, (2) automated inspection mode - where all the functions are performed by the computer and the role of the human is that of a supervisor, and (3) hybrid inspection mode - where inspection functions can be allocated to the human, the machine or both.